Theoretical Analysis of ‘Demonetisation’

The following exercise is an attempt to work out the possible consequences of the recent “demonetisation” drive in India. On 8 November 2016, the Government of India stripped ₹500 and ₹1,000 denomination currency notes of their legal tender status. The primary goals were to rid the economy of fake currency and hit out at tax evaders who had amassed their illegitimately acquired wealth in the form of high value currency notes. Even if the government’s intentions are laudable, the sudden disappearance of a substantial part of currency notes (around 86%) from the economy¹ has caught the population by surprise and unleashed an unprecedented monetary turmoil.

Rajakumar and Shetty (2016) have already presented an analysis of the demonetisation exercise. In particular, they have provided a history of demonetisation in India, comparing the current demonetisation with the one carried out in 1978. The earlier one, given the high denominations of the currency demonetised relative to those normally used for transactions, did not have any major impact on public life. Besides, the purpose of the earlier demonetisation exercise was different from the present one. Quite apart from the historical perspective, these authors come up with numbers linking currency and gross domestic product (GDP) growth. After carrying out a detailed analysis of the possible effect on the economy, the authors conclude that the present demonetisation was uncalled for given the goals of the policy.

Ashok Nag (2016) too views the historical perspective of demonetisation and questions the legality of the latest exercise. Quite apart from the rich data he presents in support of his arguments, he constructs examples to illustrate the kind of difficulties that the public and the economy will face as a result of this sudden decision.

The present exercise adds a classroom dimension, analysing the issues with the aid of an elementary macroeconomic model. It is intended to be an investigation to which even beginners studying economics will be able to relate. The arguments, therefore, do not cross the boundaries of simple diagrams and steer clear of a full-fledged model-building attempt with the help of mathematical and econometric techniques. Needless to say, simple-minded algebra can also be utilised to express the ideas developed here. But, we choose not to follow that route, in the hope of attracting as large an audience as possible.

The next section discusses how the monetary framework of the economy is affected by the decision to demonetise. The section “Goods Market Equilibrium” goes on to discuss the impact of demonetisation on non-monetary variables, such as demand for and supply of goods and services in an economy that does not engage in international trade. Next, the “IS–LM Equilibrium” presents the interactions between the monetary and non-monetary variables for the closed economy. “Open Economy Considerations” moves on to open economy implications by allowing for international transactions involving both the current and capital accounts. This is followed by “Impact on Prices and Inflation,” which discusses the possible behaviour of the general price level, which is artificially held fixed in the previous sections. The last section is titled “Concluding Remarks.”

Money Market Equilibrium

We shall assume to start with that the aggregate price level (or, the GDP deflator) P is fixed.² By definition, the nominal money supply, M, equals currency plus bank deposits, denoted CU + D.³ Suppose c is the fraction of money held as currency and (1-c) the fraction held as deposits.

Then,

M = CU + D = cM + (1 – c)M

Once again, by definition, total high-powered money (issued by the Reserve Bank of India or RBI) is H, and equals currency plus reserves, or CU + R. The institutionally specified minimum reserve deposit ratio (commonly referred to as the cash reserve ratio) is θ. To begin with, we shall assume that the banks are fully loaned up, so that R/D = θ, or, R = θD. As we shall see below, however, in the immediate short run at least, the banks may end up with excess reserves following the demonetisation move.

Prior to demonetisation then,

H = CU + R = cM + θ(1 – c)M

and

 

where   represents the so-called money multiplier.

We shall assume that c, the propensity to transact in cash, is not significantly affected by demonetisation. An important reason underlying this assumption is that the banking system is yet to penetrate vast rural areas. Besides, for whatever social and historical reasons, cash transaction is preferred by the majority in the country.

Let us now study two balance sheets, the RBI balance sheet and the aggregate balance sheet of the commercial banks. They will have roles to play in the analysis. We present simplified versions of these balance sheets in turn. To start with, consider the RBI balance sheet (Table 1).

Prior to the demonetisation announcement, the value of assets exactly matched the value of liabilities in each account. The condition needs to be satisfied post the demonetisation as well.

Thus, before demonetisation,

B_R + O = R_C + CU + R_g, ...(1)

where R_C + R_g = R = total reserves.

The government specified a time limit, say T (or, 30 December 2016 to be precise), by which the public was required to deposit ₹500 and ₹1,000 notes in its possession (to be referred to simply as “notes” hereafter) to approved authorities, in particular to the commercial banks. Apart from the time limit, there is an upper limit on the amount too that can be deposited by an individual account holder. Crossing that limit, say N, will lead to scrutiny of the income source of the deposit, to determine if taxes had been paid on the earnings. The income tax officials in charge are empowered to impose steep penalties on suspected assesses. This possibility is expected to prevent tax evaders in possession of notes in excess of N from depositing the money into their own accounts.⁴

The implications of this move can be understood by bringing in the aggregate balance sheet of the commercial banks as well (Table 2).

Like the RBI, the commercial banks’ assets must equal their liabilities. Hence, R_c + B_c + L = D.

As the public deposited notes, D rose simultaneously with R_c in the aggregate commercial banks’ account presented above. However, D fell too on account of cash withdrawals, but withdrawals were not permitted beyond an upper limit W (where W < N ). Given that W < N, the reserves R_c rose more than D in the aggregate commercial banks’ account.⁵ As a result, the effective reserve–deposit ratio rose for a while above θ, say to θ´ > θ.

The last inequality brought down the money multiplier to⁶

 

 

Subsequently, however, the government absorbed the excess liquidity from the banking system by raising the ceiling on its bond holdings under the market stabilisation scheme and this, hopefully, prevented the apprehended fall in the value of the money multiplier.⁷

What was the impact of demonetisation on H? Prior to demonetisation, the RBI’s account was balanced. Denote now the value of the notes (net of withdrawals) deposited by the public to the banks by ΔCU. As already noted, they transform into a rise in the commercial banks’ reserves to R_C + ΔR_C, where ΔR_C = ΔCU.

In the RBI, the liability entries change. Commercial banks’ reserves R_C rise to R_C + ΔR_C, whereas C changes to C – ΔCU. A certain sum V (for “vanished”) was expected not to come back to the banks, representing the government’s estimate of black money in circulation. Till date, however, V has fallen far short of the government’s expectations.⁸ The value of V will only be known later.

Assuming V > 0, however, the RBI’s liabilities add up to

R_c + ΔR_c + CU – ΔCU – V + R_g = R_c + CU + R_g – V

< B_R + O ...(2)

The last inequality in (2) follows from (1). If V > 0, the RBI’s total liabilities should fall below its assets. However, the RBI governor explained that V will continue to remain the RBI’s liability, whether or not the notes underlying it have lost legal tender status. The RBI is committed to accept them back against legal tender currency, should they be brought back. In view of this, the accounting value of the RBI’s monetary liability (H) remains unchanged. Effectively though, the monetary base will fall unless the entire V turns wholly “visible” (instead of vanishing). In what follows, we shall assume that V ≠ 0, but it could be small.⁹ Thus, despite the fact that the stock H of high powered money remains unchanged in an accounting sense, in reality, it is likely to be somewhat smaller.

Even if the money multiplier remains unchanged therefore, there will be a fall in the money supply (unless, fortuitously, V = 0). What will the money market equilibrium look like? Suppose, the money demand curve is ,

 

 

where     ^M _P  denotes “real money,” Y is the aggregate real output and i is the nominal rate of interest on government bonds. With a smaller money supply, the change in the equilibrium value of i is captured in

Figure 1 (p 69) as a rise from i₁ to i₂.

The standard macroeconomic argument underlying the interest rate rise is as follows. At the old equilibrium rate of interest, there is excess demand for money, given the fall in money supply. This leads the financial institutions, including commercial banks, to sell government bonds, lowering bond prices p_B and raising i.

The impact on the so-called LM curve is captured by Figure 2. It shifts up at each value of Y from LM₁ to LM₂.

To repeat, the extent of the shift may not be too large, given that the final value of V could turn out to be inconsequential. This concludes our discussion of the impact of demonetisation on the money market.

Goods Market Equilibrium

The goods market equilibrium is normally captured by the equation

Y = C(Y) + I(i) + G

We propose to change the consumption function in four ways. First, we make it inversely related to i_b, where i_b is the rate of interest on bank loans.
A fall in i_b is expected to increase the demand for durable consumer goods, cars being the standard example. Housing loans may not play an important role, since the government has been making repeated announcements that real estate transactions are being monitored. In the investment function too, i_b must replace i, though it is not clear how elastic I is to changes in i_b. The output level Y might also affect I, but we are ignoring this possibility without loss of generality. We assume that i_b = i + ρ, ρ > 0, to capture a risk premium associated with lending to parties other than the government.

The second variable that should enter the consumption function is an index representing ease of carrying out transactions. In our case, de-recognition of notes acts as a barrier to consumption and perhaps to investment also. The arrival of ₹2,000 notes and the delay in issuing new notes of different denominations too are constituting an impediment to transactions. Retail trade, and through backward linkages, wholesale trade as well, lose strength as a result of this ease of transaction effect. In particular, agricultural production, harvesting, etc, are affected. We shall denote this variable by e.

The third variable is consumer confidence. The government’s announcements that worse penalties are in store can dampen expenditure. This variable might impact investment also, but we will ignore that effect. Denote the confidence variable by φ. Manmohan Singh has written a most thought-provoking article on this issue.¹⁰

The fourth variable that may be introduced into the consumption function is a Patinkin (1965) type real balance effect, representing the positive impact of wealth on expenditure. Let us denote the total wealth relevant for the consumption function by W, of which V constitutes a part. Introducing these changes, the goods markets equilibrium equation is written as

Y = C (Y, i + ρ, e, φ, W) + I (i + ρ, e) + G ...(3)

Reductions in e, φ and W reduce C and I as of any given value of i. The fall in W comes about through a reduction in W to W–V. Given G, e, φ and W, equation (3) captures the standard IS curve of macroeconomic theory, relating the interest rate i to the aggregate output Y. The negative impacts of the fall in e, φ and W shift the IS curve leftwards. The larger these effects, the larger the potential fall in expenditure and output.

IS–LM Equilibrium

Let us now consider the IS–LM equilibrium prior to and following demonetisation. After demonetisation, both IS and LM shift to the left (or upwards), thus lowering Y certainly. In view of the arguments in “Money Market Equilibrium,” the LM curve may not have shifted significantly. The impact on i or i + ρ is ambiguous. Figure 3 assumes though that the rate has gone down from i₁ to i₂.^{11, 12}

A fall in the money supply alone, is normally expected to cause a rise in the sovereign rate of interest. However, the rate has fallen and the fall has obviously been caused by a sharp decline in activity in the real sector (see note 11). The IS curve has presumably shifted sharply to the left on account of the changes in e, φ, and the fall in W (causing a fall in aggregate output from Y₁ to Y₂. This lends solid theoretical support to Manmohan Singh’s view quoted (see note 10). It is the fall in Y that had brought about the fall in the rate of interest. That is, the demand curve for money shown in Figure 1 had probably moved to the left and more than reversed the initial rise in the rate of interest linked to a fall in the money supply. Notice that the Government of India was pushing the RBI for a while to reduce interest rates in the hope that such a move could increase demand and output. The demonetisation exercise has achieved the goal of a reduction in the interest rate, but for the wrong reason. It is the fall in output that has lowered the interest rate. Further, the fall in the output level could be associated with a fall in the GDP growth rate for the third quarter (October–December) and perhaps even the fourth of the present financial year.¹³

How will the scenario change beyond T? Beyond T, given the RBI announcement that its balance sheet will not be altered as of now in favour of the government, there will be no V backed rise in G as things stand now. The IS curve is therefore not expected to shift up either, unless spending sentiments change. Given the “nature” of expenditure dampening daily announcements by the government though, private sector demand is unlikely to receive a boost. Therefore, output will almost surely be slow to improve. Theory suggests then that the rate of growth of GDP for the present financial year is unlikely to reach the value projected in the Union Budget for 2016–17.

Open Economy Considerations

We have proceeded so far under closed economy assumptions. However, India is not a closed economy and it is worth our while to extend the analysis carried out so far to an open economy model. If the economy is open, then the goods market equilibrium condition changes to

Y C (.) + I(.) + G + X(Y*,  ε) – IM (Y,  ε)/ ε,

where Y* is the level of foreign incomes and ε is the real rate of exchange.¹⁴ When the domestic interest rate on government bonds falls (as of a given rate of return on, say, US bonds (say) and the expected future nominal exchange rate E^e), the domestic nominal as well as real rates of exchange are likely to fall on account of capital flight. However, the positive side of the story is that the fall in ε can improve the balance of trade and this in turn will have a positive impact on output, which could partly offset the fall in output described earlier.

How has the exchange rate behaved since 8 November? Till date (that is, 10 December), its behaviour is captured by Figure 4.¹⁵ There was a sharp drop in the nominal rate from 9 November to 21 November or so, indicating the immediate impact of the fall in the rate of return on sovereign bonds. Beyond 21 November, the nominal rate began to rise and the turnaround is being explained by most financial dailies (for instance, Goyal 2016) in terms of RBI intervention in foreign exchange markets to prevent the rupee from sliding further. The appreciation of the rupee should have reversed the positive impact of depreciation on output and to this extent the positive and negative impacts of the currency movement on output ought to have cancelled each other out, leaving our conclusion in the previous section unchanged.

Impact on Prices and Inflation

One hoped that the fall in output was not associated with a squeeze in agricultural output. However, news reports on the rural economy are depressing. Here is what LiveMint reports on 23 November:

The impact is visible in different sub-segments. Winter crops such as wheat, mustard, chickpeas are due for sowing in a fortnight. Wheat prices were already up due to low stocks and anticipated shortfall in 2015–16 output and have firmed up further as demonetisation fallout pushes traders to build more inventories. Production in 2016–17 could drop if sowed acreage (rabi) reduces for want of enough seeds on time to exploit the adequate soil moisture. Yields could fall from late sowing and subsequent exposure to rough spring weather, the lack of sufficient or timely application of fertilizers, pesticides, etc. Farm labour, vital for this period, is reported to be unpaid as farmers have no cash. Many of them are reported to be returning from some northern parts to homes in UP and Bihar. Labour shortages and wage-spikes may follow with a lag.

If the news is to be believed, then food prices can go up leading to a general rise in prices. Prices can rise on account of the exchange rate depreciation compared to the pre-demonetisation level (Figure 4) as well, since this will increase the value of imports, especially in the face of crude prices firming up in the world markets. The latest bi-monthly review of the RBI did take these inflationary possibilities into account and kept the repo rate unchanged. The signal that monetary policy was not going to ease immediately (the real money supply M/P might even decrease if P rises, given M) led to a jump in the sovereign rate too, which indicates that the exchange rate will probably not depreciate in the very near future. However, the rise in the interest rate is likely to reduce output even more than what was indicated in Figure 3, on account of an upward movement along a sticky IS curve (accompanied by a possible fall in M/P). Whether the inflation rate will be affected as well is not immediately clear, but if there is a demonetisation-generated crop failure during 2016–17 and crude prices continue to rise, then stagflation possibilities cannot be ruled out. Further, the government’s present effort to create a cashless society at short notice will not ease matters. The Indian society will take a long time yet to adjust to cashless transactions. The absence of widespread banking facilities, inadequate electrification and illiteracy stand in the way.

Concluding Remarks

Will the demonetisation exercise produce any permanent benefit? From the analysis carried, the short- to medium-run scenario does not appear to be too rosy. None of the economic variables of importance are likely to move in a healthy direction. Further, if corruption itself cannot be addressed, we may very well end up with a scenario where new black money will drive out old black money from the system.¹⁶

We may draw the readers’ attention to a piece of advice from the rational expectations school of thought in macro­economics. Even if the profession in general does not subscribe to the theoretical foundations of that school any longer, one of its policy conclusions probably cannot be ignored, especially in the context of the Indian government’s claim that in order to ensure the success of its demonetisation drive, it had decided to catch the public unawares. In other words, the government believes that its action has been successful because it took the form of a random shock to the system. It is in this context that we should take note of a well-known observation made by Robert Lucas (1997) in his Nobel lecture: “Unanticipated monetary … contractions can induce depression.”

India’s random demonetisation exercise therefore may well turn out to be a test case for Lucas’ prediction.

Notes

1 The economy in question is one, where on account of established habits, illiteracy, absence of banking facilities in vastly spread non-metropolitan areas, and a host of other reasons, cashless transactions constitute the norm rather than the exception.

2 The assumption will be relaxed later on.

3 M, in the present context represents M₁.

4 Though, the possibility of money laundering exists. A person can deposit her/his excess cash into the accounts of willing people, including poorer sections of the population. According to media reports, the near zero-balance Jan Dhan accounts are already being used for the purpose. The person who obliges is likely to be offered a price for the service, thereby creating new black money. Of course, the government has announced that it is keeping a close watch over the Jan Dhan accounts to prevent their misuse.

5 A caveat is in order here. While there is a limit on cash withdrawals, there is no limit on cheque transactions. A cheque transaction amounts to a redistribution of the aggregate reserves among banks, leaving R_C unchanged. But, cash withdrawals from deposits do change R_C.

6 According to the Business Standard (2016), “In order to absorb the surge in liquidity in (the) banking system following demonetisation of high value notes, the RBI introduced an incremental cash reserve ratio of 100 percent for the fortnight beginning Saturday.”

7 Times of India (3 December 2016) stated, “The government on Friday increased the ceiling of special bonds under the Market Stabilisation Scheme (MSS) that can be issued to mop up excess liquidity in the system to ₹6,00,000 crore, from the earlier ₹30,000 crore, in a move that will help absorb excess liquidity in the system while also helping banks to park their surplus funds and earn interest.”

8 The total estimated value of notes (in the sense the term “notes” has been defined above) in circulation was ₹14.5 lakh crore, of which ₹8.5 lakh crore was estimated to be black. That is, the government expected the value of V to b ₹8.5 lakh crore. In the bi-monthly review held on 8 December 2016, the RBI announced that ₹11.55 crore of notes have been returned to the system. This means that the actual value of V has fallen significantly short of the expected value. The government may have estimated the quantum of black money incorrectly. More likely, the method adopted to weed out black money lacked efficiency. A good deal of black money appears to have been laundered through corrupt use of bank accounts.

9 There was speculation that V will be credited to the government’s account in the RBI in the form of dividend receipts. According to the budget 2016–17, the government’s dividend
receipts from RBI and nationalised banks for the 2015–16 was ₹69,897 crore. The RBI’s dividend payment to the government for 2016–17 will soon be known, but from the announcements made so far, the dividend will not be linked to V. In any case, the governor of RBI has explained that there is no asset liability mismatch for RBI on account of V, whether V turns visible or not.

10 To quote Manmohan Singh (2016), “Consumer confidence is an important economic variable in a nation’s growth prospects. It is now evident that this sudden overnight ban on currency has dented the confidence of hundreds and millions of Indian consumers, which can have serious economic ramifications. The scars of an overnight depletion of the honest wealth of a vast majority of Indians combined with their ordeal of rationed access to new currency will be too deep to heal quickly. This can have ripple effects on GDP growth and job creation. It is my humble opinion that we as a nation should brace ourselves for a tough period over the coming months, needlessly so.”

11 This assumption is motivated by the data source in the following link: http://in.investing.com/rates-bonds/india-10-year-bond-yield-streaming-chart. This shows that the rate of return on 10-year bonds had fallen during the period 8 November through 6 December from around 6.80% to around 6.20%. Following the bi-monthly review of 8 December of course, when RBI announced its intention of sticking to a tight monetary policy, the rate shot up to 6.44% on 9 December.

12 A likely consequence of the fall in the sovereign rate may be a fall in the interest rates on savings as well as term deposits of commercial banks.

13 By definition, the quarterly growth rate compares the present quarter to the same quarter in the previous year.

14 The real rate of exchange is defined here as the price of Indian goods in units of US goods. This is a simplification, since, strictly speaking, India has other trading partners as well. However, a good part of our trade is quoted in terms of the US dollar. Hence, we are ignoring other currencies. The nominal exchange rate E equals the dollar price of a rupee. Thus, depreciation/appreciation of the rupee amounts to a fall/rise in E. Moreover, ε, where P* is the given US price level.

15 As with note 11, one will be able to follow
future movements of the exchange rate in the following link. Data source: http://www.x-rates.com/graph/?from=INR&to=USD&amount=1.

16 And it is best to point out that the tools of analysis employed by this note are not relevant for studying the long-run. Quite apart from the fact that long-term expectations will play a role, one needs to study long-term developments with the aid of properly specified growth models. Consequently, these short- and medium-run methods have no implications on the claim that short-run pains will be compensated by long-run gains.

Reference

Business Standard (2016): “RBI Introduces Incremental CRR to Manage Excess Liquidity,” 26 November.

Goyal, Kartik (2016): “RBI Intervention Brakes Rupee Near Record Low Amid $2.7 Billion in Outflows,” Economic Times, 23 November, http://economictimes.indiatimes.com/markets/forex/rbi-intervention-brakes-rupee-near-record-low-amid-2-7-billion-in-outflows/articleshow/55577605.cms.

Lucas, Robert (1997): “Monetary Neutrality,” Nobel Lectures, Economics, 1991–95, Torsten Persson (ed), Singapore: World Scientific Publishing Company.

Nag, Ashok K (2016): “Lost Due to Demonetisation,” Economic & Political Weekly, Vol 51, No 48,
pp 18–21.

Patinkin, Don (1965): Money, Interest and Prices, New York: Harper and Row.

Rajakumar, J Dennis and S L Shetty (2016): “Demonetisation: 1978, the Present and the Aftermath,” Economic & Political Weekly, Vol 51, No 48, pp 13–17.

Singh, Manmohan (2016): “Making of a Mammoth Tragedy,” Hindu, 9 December.